Efficent Lane Change Path Planning based on Quintic spline for Autonomous Vehicles

This paper proposes a lane change path planning method for Autonomous Vehicles (AVs) driving in structured environments. Due to the complexity of dynamically changing structured environment, the path planner generally requires to generate a safe and smooth path in a short time when changing lanes. Therefore, we propose an efficient lane change path planning method, which can take into account both the curvature and heading angle of the planned path, ensuring the safety and stability of the lane change maneuver. It builds upon existing method using quintic spline with second-order geometric (G
²
-) continuity in a Frenet coordinate system. By calculating the optimal prediction time interval of the lane change maneuver, the heading angle was kept within a safe range. In order to improve the real-time performance, we use road prior information to optimize the selection of sampling points, effectively reducing the number of candidate paths. In addition, beyond the sampling points, the problem of continuous lane change is solved by adding mid-adjust point and target point, so that AVs have the ability to complete complex movements. In an urban environment, the proposed method is tested on our experimental autonomous vehicle equipped with a modular ROS architecture. Experimental results show that the proposed algorithm can generate complex motion paths efficiently, and the curvature and heading angle of the planned path are continuous, which effectively improves the efficiency and stability of path planning.